Abstract
Hand-harvesting parameters and postharvest marketability attributes of fresh-market blackberries (Rubus L. subgenus Rubus Watson) were characterized to develop a prototype for a soft-robotic gripper for robotic harvesting. A custom-made, force-sensing apparatus attached to the thumb and fingers of a person hand-harvesting blackberries was developed to quantify forces used to harvest and to identify appendages for harvesting. Four cultivars of blackberries grown in Arkansas were harvested at optimal ripeness and stored at 2 °C for 21 days to determine the impact on marketability attributes (leakage, decay, and red drupelet reversion). The forces during harvest imparted by the thumb and middle finger were greatest (0.77 N and 0.37 N, respectively), whereas the index and ring fingers used lower forces (0.16 N and 0.06 N, respectively), primarily to stabilize the blackberry. The forces applied to grab, stabilize, and harvest blackberries caused minimal marketability damage (leakage, <10%; decay, <2%; and red drupelet reversion, <8%) after postharvest storage. This project quantified harvest and postharvest parameters, allowing data-driven design of a three-prong soft-robotic gripper for harvest of fresh-market blackberries.
Fresh-market blackberries (Rubus L. subgenus Rubus Watson) are hand-picked to maintain quality from harvest to consumption. Mechanical harvesting of processing blackberries decreases labor costs and harvest time (Takeda and Peterson, 1999). However, the delicate nature of fresh-market blackberries limits mechanical harvesting options and hampers industry expansion. As an alternative to harvesting fruit with rigid robotics or mechanical systems, soft robotics use compliant structural materials, such as rubber and silicone, that enable grippers to grasp objects while reducing cost and complexity. This technology allows task versatility ideal for grasping and manipulating delicate objects (Shintake et al., 2018) with dexterity that emulates human hand grasping (Venter and Dirven, 2017). Although soft robots are a potential harvesting solution for fresh-market blackberries, little is known about the forces associated with hand-harvesting blackberries. Thus, physical, composition, and marketability attributes of blackberries, along with forces and appendages used for hand-harvesting blackberries, were evaluated to provide guidance for developing a soft-robotic gripper.
Materials and Methods
Cultivars and harvest.
A custom-made force-sensing apparatus attached to the thumb and fingers of a person was developed to quantify forces for harvesting blackberries and to identify appendages essential for harvesting (Fig. 1). The apparatus was designed with resistive force sensors (FlexiForce A301; Tekscan, South Boston, MA) placed in holes cut into silicone finger sleeves (sensors extended 0.5 mm beyond the sleeve). The silicone finger sleeves were positioned on the thumb and three fingers (index, middle, and ring) of the right hand and were oriented to maximize contact with berry surfaces during harvesting. Sensor validation and calibration ensured accurate force measurements. Voltage data were measured via a single power source noninverting op-amp circuit sent through Bluetooth to MATLAB (MathWorks, Inc. Natick, MA), and converted to force values. Data recording and processing were conducted in a portable, water-resistant case housed in a backpack.
Custom-made force-sensing apparatus with sensors on the thumb and three fingers (left) placed on the hand for harvesting (right) fresh-market blackberries (Arkansas, 2020).
Citation: HortScience 57, 5; 10.21273/HORTSCI16487-22
Four blackberry cultivars—Natchez, Osage, Prime-Ark® Traveler, and Sweet-Ark® Caddo—were harvested in June and July 2020 from commercial growers in Arkansas. About 2 kg of each cultivar were harvested into 170-g vented clamshells. For each cultivar, 240 berries were harvested (20 berries/clamshell for four storage times in triplicate). After harvest, the clamshells of blackberries were transported to the University of Arkansas System Division of Agriculture Food Science Department, Fayetteville, AR, for analysis of physical, composition, and marketability attributes at harvest (day 0), and marketability attributes during storage (0, 7, 14, and 21 d) at 2 °C. Only 0- and 21-d postharvest storage data are presented for physical, composition, and marketability attributes.
Physical attribute analysis.
The physical attributes were evaluated using five berries/cultivar and replication. Each berry was weighed (in grams) using a precision digital scale, then length and width (both in millimeters) were measured using digital calipers. The firmness of each berry was measured using a Stable Micro Systems TA.TX. XT plus Texture Analyzer (Texture Technologies Corporation, Hamilton, MA). Each berry was placed horizontally on a flat surface and compressed with a 7.6-cm-diameter cylindrical probe using a trigger force of 0.02 N. The force needed to compress the berry was, as mentioned, measured in Newtons. The berries were frozen at –10 °C for composition analysis.
Composition analysis.
Composition of the juice from five berries/cultivar and replication were measured for soluble solids, pH, and titratable acidity. The five berries were thawed and squeezed through cheesecloth to extract juice for analysis. Soluble solids (measured as a percentage) of the juice were measured using an Abbe Mark II refractometer (Bausch and Lomb, Scientific Instrument, Keene, NH). pH and titratable acidity (percent citric acid) were measured using a Metrohm AG 862 Compact Titrosampler (Herisau, Switzerland).
Marketability analysis.
Marketability attributes evaluated included decay, leakiness, and red drupelet reversion (RDR) at 0 and 21 d at 2 °C. The decay (visible mold/rot on berry), leakiness (rolling berry on white paper towel), and RDR (one or more red drupelets on berry) of each blackberry were evaluated. The percentage of each attribute for the blackberries in clamshells was calculated as (number of impacted berries/number of total berries) × 100.
Statistical design and analysis.
Four cultivars were evaluated with three replications at harvest and after storage. Data were analyzed by analysis of variance using JMP® (version 16.0; SAS Institute Inc., Cary, NC). Means with different letters for each attribute within effects were significantly different (P < 0.05) using Tukey’s honestly significant difference test.
Results and Discussion
For physical attributes, berry weight, width, and firmness were impacted by cultivar but not length (26.87 mm) (Table 1). Sweet-Ark® Caddo (8.10 g) and ‘Natchez’ (6.26 g) had the largest berries; Prime-Ark® Traveler (4.88 g) and ‘Osage’ (3.76 g) had the smallest. ‘Osage’ (19.05 mm) was narrower, and Sweet-Ark® Caddo (8.78 N) was firmer than the other cultivars. For composition attributes, pH and titratable acidity were impacted by cultivar but not soluble solids (12.16%). ‘Osage’ and Prime-Ark® Traveler (3.39 and 3.33, respectively) were higher in pH than ‘Natchez’ (3.22). ‘Natchez’ (1.40%) had the highest titratable acidity, followed by Sweet-Ark® Caddo (1.17%), ‘Osage’ (1.12%), and Prime-Ark® Traveler (1.07%). Previous research on Arkansas-grown blackberries found similar ranges in fruit size, firmness, and composition (Felts et al., 2020; Segantini et al., 2017).
Physical attributes of fresh-market blackberries and force-to-harvest blackberries using a custom-made apparatus with sensors on the thumb and three fingers placed on the hand (Arkansas, 2020).
Regardless of cultivar, the thumb applied the greatest force (0.77 N) to harvest blackberries, followed by the middle (0.37 N), index (0.16 N), and ring (0.06 N) fingers. The force to harvest blackberries using the thumb (0.51–1.18 N), middle (0.31–0.49 N), index (0.09–0.27 N), and ring (0.01–0.15 N) fingers was different for each cultivar (Table 1). Sweet-Ark® Caddo (1.18 N) had a greater force on the thumb than the other cultivars. ‘Natchez’ had the greatest force on the index (0.27 N) and middle (0.49 N) fingers. In general, berry weight and firmness were related to the force needed to harvest. Sweet-Ark® Caddo—the largest, firmest berry—had the greatest force to harvest on the thumb, whereas ‘Osage’ (0.65 N) and Prime-Ark® Traveler (0.51 N)—the smallest berries—had the least force to harvest. The thumb and middle finger were primarily force applicators, whereas the index and ring fingers stabilized the berry. The underuse of the ring finger to harvest blackberries justified the design of a three-finger/prong robotic gripper.
Cultivar did not impact marketability attributes at harvest (<27% leakage, 0% decay, and <2% RDR) or at 21 d of storage at 2 °C (<10% leakage, <2% decay, and <8% RDR). These results show that although the force to harvest blackberries differed, the forces did not impact marketability attributes detrimentally. Thus, a soft-robotic gripper should be designed to use similar force to ensure marketable blackberries.
Conclusion
A force-sensing apparatus for hand-harvesting blackberries was developed to quantify forces applied by the thumb and fingers during harvesting, and to identify essential appendages. Across cultivars, the force used by the thumb and middle finger were greatest, whereas the index and ring fingers used lower forces to harvest blackberries, acting primarily as berry stabilizers. In addition, the forces applied to grab, stabilize, and harvest the blackberries caused minimal damage after postharvest storage. This project determined harvest and postharvest marketability attributes that were used to develop a novel, three-prong soft-robotic gripper for robotic harvesting of fresh-market blackberries (Gunderman et al., 2021, 2022).
Literature Cited
Felts, M., Threlfall, R.T., Clark, J.R. & Worthington, M.L. 2020 Effects of harvest time (7:00 AM and 12:00 PM) on postharvest quality of Arkansas fresh-market blackberries Acta Hort. 1277 477 486 https://doi.org/10.17660/ActaHortic.2020.1277.68
Gunderman, A.L., Collins, J., Myer, A., Threlfall, R. & Chen, Y. 2021 Tendon-driven soft robotic gripper for berry harvesting arXiv https://doi.org/10.48550/arXiv.2103.04270
Gunderman, A.L., Collins, J.A., Myers, A.L., Threlfall, R.T. & Chen, Y. 2022 Tendon-driven soft robotic gripper for blackberry harvesting IEEE Robot. Autom. Lett. 7 2 2652 2658 https://doi.org/10.1109/LRA.2022.3143891
Segantini, D.M., Threlfall, R.T., Clark, J.R., Brownmiller, C.R., Howard, L.R. & Lawless, L.J.R. 2017 Changes in fresh-market and sensory attributes of blackberry genotypes after postharvest storage J. Berry Res. 7 2 129 145 https://doi.org/10.3233/JBR-170153
Shintake, J., Cacucciolo, V., Floreano, D. & Shea, H. 2018 Soft robotic grippers Adv. Mater. 30 29 10707035 https://doi.org/10.1002/adma.201707035
Takeda, F. & Peterson, D. 1999 Considerations for machine harvesting fresh-market eastern thornless blackberries: Trellis design, cane training systems, and mechanical harvester developments HortTechnology 9 16 21 https://doi.org/10.21273/HORTTECH.9.1.16
Venter, D. & Dirven, S. 2017 Self-morphing soft-robotic gripper for handling and manipulation of delicate produce in horticultural applications 24th International Conference on Mechatronics and Machine Vision in Practice Auckland, New Zealand 1 6 https://doi.org/10.1109/M2VIP. 2017.8211516